346 may be used to communicate information between the various components of the video encoder.

An input digital video sequence is provided to the coding control component 340. The coding control component 340 sequences the various operations of the video encoder, i.e., the coding control component 340 runs the main control loop for video encoding. For example, the coding control component 340 performs processing on the input video sequence that is to be done at the picture level, such as determining the coding type (I, P, or B) of a picture based on a high level coding structure, e.g., IPPP, IBBP, hierarchical-B, and dividing a picture into LCUs for further processing.

In addition, for pipelined architectures in which multiple LCUs may be processed concurrently in different components of the LCU processing, the coding control component 340 controls the processing of the LCUs by various components of the LCU processing in a pipeline fashion. For example, in many embedded systems supporting video processing, there may be one master processor and one or more slave processing modules, e.g., hardware accelerators. The master processor operates as the coding control component and runs the main control loop for video encoding, and the slave processing modules are employed to off load certain compute-intensive tasks of video encoding such as motion estimation, motion compensation, intra prediction mode estimation, transformation and quantization, entropy coding, and loop filtering. The slave processing modules are controlled in a pipeline fashion by the master processor such that the slave processing modules operate on different LCUs of a picture at any given time. That is, the slave processing modules are executed in parallel, each processing its respective LCU while data movement from one processor to another is serial.

When weighted prediction is enabled for a picture, i.e., weighted prediction is enabled for one or both of any P slices or B slices in the picture, the weighted prediction analysis component 344 determines whether or not luminance and/or chrominance weighted prediction is to be used for each applicable reference picture and sets the corresponding weight flags as appropriate. The applicable reference pictures are the list 0 reference pictures if weighed prediction for P slices is enabled, and the list 0 and list 1 reference pictures if weighted prediction for B slices is enabled. Further, if weighted prediction is to be used for luminance and/or chrominance for a reference picture, the weighted prediction analysis component 344 determines the weighting factors to be used. Note that enabling weighted prediction for P slices and/or B slices is an encoder option and an encoder may use any suitable technique for deciding whether or not weighted prediction for one or both slice types is to be enabled.

Rather than allowing weighted prediction to possibly be enabled for both luminance and chrominance for all applicable reference pictures as in the prior art, the weighted prediction analysis component 344 limits the total number of weighting factors for a picture to be less to be less than or equal to a predetermined threshold number of weighting factors, e.g., 24, specified by HEVC. The operation of the weighted prediction analysis component 344 is explained in more detail in reference to the methods of FIGS. 5 and 6 below. The weighted prediction analysis component 344 provides the luminance and chrominance weight flags, and the associated weighting factors to the coding control component 340 for communication to various components of the LCU processing component 342 as needed.

FIG. 3B is a block diagram of the LCU processing component 342. The LCU processing component 342 receives LCUs 300 of the input video sequence from the coding control component and encodes the LCUs 300 under the control of the coding control component 340 to generate the compressed video stream. The LCUs 300 in each picture are processed in row order. The LCUs 300 from the coding control component are provided as one input of a motion estimation component (ME) 320, as one input of an intra-prediction estimation component (IPE) 324, and to a positive input of a combiner 302 (e.g., adder or subtractor or the like). Further, although not specifically shown, the prediction mode of each picture as selected by the coding control component 340 is provided to a mode decision component 328 and the entropy coding component 336.

The storage component 318 provides reference data to the motion estimation component 320 and to the motion compensation component 322. The reference data may include one or more previously encoded and decoded pictures, i.e., reference pictures. Both list 0 and list 1 reference pictures may be stored by the storage component 318.

The motion estimation component 320 provides motion data information to the motion compensation component 322 and the entropy coding component 336. More specifically, the motion estimation component 320 performs tests on CUs in an LCU based on multiple inter-prediction modes (e.g., skip mode, merge mode, and normal or direct inter-prediction), PU sizes, and TU sizes using reference picture data from storage 318 to choose the best CU partitioning, PU/TU partitioning, inter-prediction modes, motion vectors, etc. based on coding cost, e.g., a rate distortion coding cost. To perform the tests, the motion estimation component 320 may divide an LCU into CUs according to the maximum hierarchical depth of the quadtree, and divide each CU into PUs according to the unit sizes of the inter-prediction modes and into TUs according to the transform unit sizes, and calculate the coding costs for each PU size, prediction mode, and transform unit size for each CU. If weighted prediction is enabled for the current picture, any weights and offsets for reference pictures generated by the weighted prediction analysis component 344 are applied as appropriate. For example, the size of prediction error between the current PU and the reference PU may be computed as the sum of absolute difference between the current PU and the weighted reference PU. The motion estimation component 320 provides the motion vector (MV) or vectors and the prediction mode for each PU in the selected CU partitioning to the motion compensation component (MC) 322.

The motion compensation component 322 receives the selected inter-prediction mode and mode-related information from the motion estimation component 320 and generates the inter-predicted CUs. If weighted prediction is enabled for the current picture and for the particular reference picture or pictures used for inter-prediction, the weights and offsets for the reference picture or pictures generated by the weighted prediction analysis component 344 are applied in generating the inter-predicted CU. The inter-predicted CUs are provided to the mode decision component 328 along with the selected inter-prediction modes for the inter-predicted PUs and corresponding TU sizes for the selected CU/PU/TU partitioning. The coding costs of the inter-predicted CUs are also provided to the mode decision component 328.

The intra-prediction estimation component 324 (IPE) performs intra-prediction estimation in which tests on CUs in an LCU based on multiple intra-prediction modes, PU sizes, and TU sizes are performed using reconstructed data from previously encoded neighboring CUs stored in a buffer (not shown) to choose the best CU partitioning, PU/TU partitioning, and intra-prediction modes based on coding cost, e.g., a rate distortion coding cost. To perform the tests, the intra-prediction estimation component 324 may divide an LCU into CUs according to the maximum hierarchical depth of the quadtree, and divide each CU into PUs according to the unit sizes of the intra-prediction modes and into TUs according to the transform unit sizes, and calculate the coding costs for each PU size, prediction mode, and transform unit size for each PU. The intra-prediction estimation component 324 provides the selected intra-prediction modes for the PUs, and the corresponding TU sizes for the selected CU partitioning to the intra-prediction component (IP) 326. The coding costs of the intra-predicted CUs are also provided to the intra-prediction component 326.

The intra-prediction component 326 (IP) receives intra-prediction information, e.g., the selected mode or modes for the PU(s), the PU size, etc., from the intra-prediction estimation component 324 and generates the intra-predicted CUs. The intra-predicted CUs are provided to the mode decision component 328 along with the selected intra-prediction modes for the intra-predicted PUs and corresponding TU sizes for the selected CU/PU/TU partitioning. The coding costs of the intra-predicted CUs are also provided to the mode decision component 328.

The mode decision component 328 selects between intra-prediction of a CU and inter-prediction of a CU based on the intra-prediction coding cost of the CU from the intra-prediction component 326, the inter-prediction coding cost of the CU from the motion compensation component 322, and the picture prediction mode provided by the coding control component. Based on the decision as to whether a CU is to be intra- or inter-coded, the intra-predicted PUs or inter-predicted PUs are selected. The selected CU/PU/TU partitioning with corresponding modes and other mode related prediction data (if any) such as motion vector(s) and reference picture index (indices), are provided to the entropy coding component 336.

The output of the mode decision component 328, i.e., the predicted PUs, is provided to a negative input of the combiner 302 and to the combiner 338. The associated transform unit size is also provided to the transform component 304. The combiner 302 subtracts a predicted PU from the original PU. Each resulting residual PU is a set of pixel difference values that quantify differences between pixel values of the original PU and the predicted PU. The residual blocks of all the PUs of a CU form a residual CU for further processing.

The transform component 304 performs block transforms on the residual CUs to convert the residual pixel values to transform coefficients and provides the transform coefficients to a quantize component 306. More specifically, the transform component 304 receives the transform unit sizes for the residual CU and applies transforms of the specified sizes to the CU to generate transform coefficients. Further, the quantize component 306 quantizes the transform coefficients based on quantization parameters (QPs) and quantization matrices provided by the coding control component and the transform sizes and provides the quantized transform coefficients to the entropy coding component 336 for coding in the bit stream.

The entropy coding component 336 entropy encodes the relevant data, i.e., syntax elements, output by the various encoding components and the coding control component using context-adaptive binary arithmetic coding (CABAC) to generate the compressed video bit stream. Among the syntax elements that are encoded are picture parameter sets, slice headers, flags indicating the CU/PU/TU partitioning of an LCU, the prediction modes for the CUS, and the quantized transform coefficients for the CUS. The entropy coding component 336 also entropy encodes relevant data from the in-loop filters, such as the SAO parameters. Further, the entropy coding component 336 encodes weighted prediction parameters in slice headers of P or B slices when weighted prediction is enabled for P slices and/or B slices in a current picture. Theses parameters include the weight flags and weighting factors generated by the weighted prediction analysis component 344. A method for signaling the weight flags and weighting factors in a slice header that may be used by the entropy coding component 336 is described below in reference to FIG. 7.

The LCU processing includes an embedded decoder. As any compliant decoder is expected to reconstruct an image from a compressed bit stream, the embedded decoder provides the same utility to the video encoder. Knowledge of the reconstructed input allows the video encoder to transmit the appropriate residual energy to compose subsequent pictures.

The quantized transform coefficients for each CU are provided to an inverse quantize component (IQ) 312, which outputs a reconstructed version of the transform result from the transform component 304. The dequantized transform coefficients are provided to the inverse transform component (IDCT) 314, which outputs estimated residual information representing a reconstructed version of a residual CU. The inverse transform component 314 receives the transform unit size used to generate the transform coefficients and applies inverse transform(s) of the specified size to the transform coefficients to reconstruct the residual values. The reconstructed residual CU is provided to the combiner 338.

The combiner 338 adds the original predicted CU to the residual CU to generate a reconstructed CU, which becomes part of reconstructed picture data. The reconstructed picture data is stored in a buffer (not shown) for use by the intra-prediction estimation component 324.

Various in-loop filters may be applied to the reconstructed picture data to improve the quality of the reference picture data used for encoding/decoding of subsequent pictures. The in-loop filters may include a deblocking filter 330, a sample adaptive offset filter (SAO) 332, and an adaptive loop filter (ALF) 334. The in-loop filters 330, 332, 334 are applied to each reconstructed LCU in the picture and the final filtered reference picture data is provided to the storage component 318. In some embodiments, the ALF component 334 is not present.

FIG. 4 is a block diagram of an example video decoder providing decoding of video sequences encoded using weighted prediction. The video decoder operates to reverse the encoding operations, i.e., entropy coding, quantization, transformation, and prediction, performed by the video encoder of FIGS. 3A and 3B to regenerate the pictures of the original video sequence. In view of the above description of a video encoder, one of ordinary skill in the art will understand the functionality of components of the video decoder without detailed explanation.

The entropy decoding component 400 receives an entropy encoded (compressed) video bit stream and reverses the entropy encoding using CABAC decoding to recover the encoded syntax elements, e.g., CU, PU, and TU structures of LCUs, quantized transform coefficients for CUs, motion vectors, prediction modes, weighted prediction parameters, SAO parameters, etc. The decoded syntax elements are passed to the various components of the decoder as needed. For example, decoded prediction modes are provided to the intra-prediction component (IP) 414 or motion compensation component (MC) 410. If the decoded prediction mode is an inter-prediction mode, the entropy decoder 400 reconstructs the motion vector(s) as needed and provides the motion vector(s) to the motion compensation component 410.

The entropy decoding component 400 decodes flags in a picture parameter set that indicate whether or not weighted prediction is enabled for P and/or B slices in the picture. If weighted prediction is enabled for P slices, for each P slice in the picture, the entropy decoding component 400 decodes weight flags and associated weighting factors (if any) for list 0 reference pictures from the slice header, and provides these parameters to the motion compensation component 410. If weighted prediction is enabled for B slices, for each B slice in the picture, the entropy decoding component 400 decodes weight flags and associated weighting factors (if any) for list 0 and list 1 reference pictures from the slice header, and provides these parameters to the motion compensation component 410. A method for decoding weight flags and associated weighting factors that may be performed by the entropy decoding component 400 is described herein in reference to FIG. 8.

The inverse quantize component (IQ) 402 de-quantizes the quantized transform coefficients of the CUs. The inverse transform component 404 transforms the frequency domain data from the inverse quantize component 402 back to the residual CUs. That is, the inverse transform component 404 applies an inverse unit transform, i.e., the inverse of the unit transform used for encoding, to the de-quantized residual coefficients to produce reconstructed residual values of the CUS.

A residual CU supplies one input of the addition component 406. The other input of the addition component 406 comes from the mode switch 408. When an inter-prediction mode is signaled in the encoded video stream, the mode switch 408 selects predicted PUs from the motion compensation component 410 and when an intra-prediction mode is signaled, the mode switch selects predicted PUs from the intra-prediction component 414.

The motion compensation component 410 receives reference data from the storage component 412 and applies the motion compensation computed by the encoder and transmitted in the encoded video bit stream to the reference data to generate a predicted PU. That is, the motion compensation component 410 uses the motion vector(s) from the entropy decoder 400 and the reference data to generate a predicted PU. Further, if weighted prediction is enabled for P and/or B slices of the current picture, the motion compensation component 410 applies weights and offsets to the reference data as appropriate during the generation of the predicted PU.

The intra-prediction component 414 receives reconstructed samples from previously reconstructed PUs of a current picture from the storage component 412 and performs the intra-prediction computed by the encoder as signaled by an intra-prediction mode transmitted in the encoded video bit stream using the reconstructed samples as needed to generate a predicted PU.

The addition component 406 generates a reconstructed CU by adding the predicted PUs selected by the mode switch 408 and the residual CU. The output of the addition component 406, i.e., the reconstructed CUs, is stored in the storage component 412 for use by the intra-prediction component 414.

In-loop filters may be applied to reconstructed picture data to improve the quality of the decoded pictures and the quality of the reference picture data used for decoding of subsequent pictures. The applied in-loop filters are the same as those of the encoder, i.e., a deblocking filter 416, a sample adaptive offset filter (SAO) 418, and an adaptive loop filter (ALF) 420. The in-loop filters may be applied on an LCU-by-LCU basis and the final filtered reference picture data is provided to the storage component 412. In some embodiments, the ALF component 420 is not present.

FIG. 5 is a flow diagram of a method for determining weighted prediction parameters for a picture that may be used, for example, by the weighted prediction analysis component 344 of the encoder of FIGS. 3A and 3B. This method assumes that weighted prediction has been enabled for P and/or B slices in the picture.

Initially, weight flags and weighting factors (weights and offsets) for weighted prediction are determined 500. If weighted prediction is enabled for P slices, luminance and chrominance weighting flag values, and weighting factors (if weighted prediction is enabled for a reference picture) are determined for list 0 reference pictures. If weighted prediction is enabled for B slices, luminance and chrominance weighting flag values, and weighting factors (if weighted prediction is enabled for a reference picture) are determined for both list 0 reference pictures and list 1 reference pictures. A method for determining weight flags and weighting factors for a picture that may be used in some embodiments is described herein in reference to FIG. 6.

After the weight flag values are determined, the total number of weighting factors is computed 502. The total number of weighting factors may be computed, for example, as the sum of the number of enabled luminance weight flags and 2× the number of enabled chrominance weight flags. More specifically, if weighted prediction is enabled for P slices, the total number of weighting factors for the list 0 reference pictures may be computed as the sum of the number of enabled list 0 luminance weight flags and 2× the number of enabled list 0 chrominance weight flags. If weighted prediction is enabled for B slices, the total number of weighting factors for both the list 0 and the list 1 reference pictures may be computed as the sum of the number of enabled list 0 luminance weight flags, 2× the number of enabled list 0 chrominance weight flags, the number of enabled list 1 luminance weight flags, and 2× the number of enabled list 1 chrominance weight flags. A weight flag is enabled if it is set to indicate that weighted prediction is to be used for the associated color component(s) of a reference picture. Further, if a luminance weight flag is enabled, there are luminance weighting factors for the associated reference picture; otherwise, there are no luminance weighting factors for the associated reference picture. If a chrominance weight flag is enabled, there are also Cb and Cr weighting factors for the associated reference picture; otherwise, there are no chrominance weighting factors for the associated reference picture.

The number of weighting factors is then compared 504 to a predetermined threshold that indicates the maximum number of weighting factors allowed. The value of this threshold is set by HEVC and may be any suitable value. In some embodiments, the value of the threshold is 24.

If the total number of weighting factors does not exceed the threshold, the method terminates. If the total number of weighting factors exceeds the threshold, the number of weighting factors is reduced 506 by selectively disabling some of the enabled weight flags and discarding the associated weighting factors. Any suitable technique may be used to determine which weighting factors are to be discarded. For example, some or all of any chrominance weighting factors may be discarded and the associated enabled chrominance weight flags disabled to reduce the number of weighting factors. In another example, some of the luminance and chrominance weighting factors for reference pictures of higher reference picture index values may be discarded and their associated weight flags disabled to reduce the number of weighting factors. After the number of weighting factors is reduced to at or below the threshold, the method terminates.

FIG. 6 is a flow diagram of a method for determining weight flag values and weighting factors for a picture. If weighted prediction is enabled for P slices, the method is executed twice, once to determine luminance weight flag values and weighting factors for list 0 reference pictures and once to determine chrominance weight flag values and weighting factors for list 0 reference pictures. If weighted prediction is enabled for B slices, the method is executed four times, once to determine luminance weight flag values and weighting factors for list 0 reference pictures, once to determine chrominance weight flag values and weighting factors for list 0 reference pictures, once to determine luminance weight flag values and weighting factors for list 1 reference pictures, and once to determine chrominance weight flag values and weighting factors for list 1 reference pictures.

Initially, the mean and variance of the picture are computed 600 for use in determining the weighting factors. If the method is being performed to determine luminance weight flag values and weighting factors, the mean and variance of the luminance component of the picture are computed. If the method is being performed to determine chrominance weight flag values and weighting factors, the mean and variance of each of the Cb and Cr components of the picture are computed.

An analysis is then performed to determine whether or not weighted prediction is to be used 602 for an initial reference picture in the reference picture list being processed. Any suitable analysis technique may be used. For example, a region based approach to detecting changes in global brightness as described in Kwon may be used. If the result of the analysis is that weighted prediction is not to be used, the weight flag for the reference picture is set 612 to indicate that weighted prediction is not enabled for the reference picture, and processing continues with the next reference picture in the reference picture list, if any 614.

If the result of the analysis is that weighted prediction is to be used, the weight flag for the reference picture is set 604 to indicate that weighted prediction is enabled for the reference picture. The mean and variance of the reference picture are also computed 606 for use in determining the weighting factors. If the method is being performed to determine luminance weight flag values and weighting factors, the mean and variance of the luminance of the reference picture are computed. If the method is being performed to determine chrominance weight flag values and weighting factors, the mean and variance of each of Cb and Cr of the reference picture are computed.

Weighting factors are then computed for the reference picture. In general, the weight W and offset O for a color component of a reference picture may be computed as

W
=

(

V
⁡
(
curr
)

/

V
⁡
(
ref
)

)

⁢

<<
shift

O
=

M
⁡
(
curr
)

-

(

(

W
*

M
⁡
(
ref
)

+

2

(

shift
-
1

)

)

>>
shift

)

- where V(curr) is the variance of the color component for the picture, M(curr) is the mean of the color component for the picture, V(ref) is the variance of the color component for the reference picture, M(ref) is the mean of the color component for the reference picture, and shift is the base 2 logarithm of the denominator for all list 0 and list 1 weighting factors in the reference picture list. Note that luminance and chrominance color components may use different denominators which are signaled in bit stream. If the method is being performed to determine chrominance weight flag values and weighting factors, weighted factors for both Cb and Cr are computed. After the weighting factors are computed, processing continues with the next reference picture, if any 614.

FIG. 7 is a flow diagram of a method for signaling weight flags and weighting factors for a reference picture list in a slice header that may be used, for example, by the entropy coding component 336 of the encoder of FIGS. 3A and 3B. If weighted prediction is enabled for P slices, the method is executed one time to signal the weight flags and weighting factors for the list 0 reference pictures. If weighted prediction is enabled for B slices, the method is executed twice, once to signal the weight flags and weighting factors for the list 0 reference pictures and once to signal the weight flags and weighting factors for the list 1 reference pictures.

First, the luminance weight flags for the reference picture list are signaled 700 in the compressed bit stream. The chrominance weight flags for the reference picture list are then signaled 702 in the compressed bit stream immediately following the luminance weight flags. Then, the luminance weighting factors corresponding to any enabled luminance weight flags are signaled 704 in the compressed bit stream. Finally, the chrominance weighting factors corresponding to any enabled chrominance weight flags are signaled 706 in the compressed bit stream.

FIG. 8 is a flow diagram of method for decoding weight flags and associated weighting factors in a slice header that may be performed, for example, by the entropy decoding component 400 of the decoder of FIG. 4. Initially, a check is made to determine if weighted prediction (WP) is enabled 800 for P or B slices in the current picture. As previously described, such enabled is signaled in the bit stream in the picture parameter set. If weighted prediction is not enabled, the method terminates.

If weighted prediction is enabled for P or B slices, luminance and chrominance weight flags for list 0 reference pictures are decoded 802 from the slice header. More specifically, luminance weight flags are decoded, followed by chrominance weight flags. Weighting factors for any enabled weight flags for the list 0 reference pictures are then decoded 804 from the slice header. More specifically, luminance weighting factors are decoded, followed by chrominance weighting factors.

If the slice type is B and weighted prediction is enabled for B slices 806, then luminance and chrominance weight flags for list 1 reference pictures are decoded